Skip to content

Latest commit

 

History

History
328 lines (244 loc) · 10.7 KB

tidb-control.md

File metadata and controls

328 lines (244 loc) · 10.7 KB
Raw
title summary aliases
TiDB Control User Guide
Use TiDB Control to obtain TiDB status information for debugging.
/docs/dev/tidb-control/
/docs/dev/reference/tools/tidb-control/

TiDB Control User Guide

TiDB Control is a command-line tool of TiDB, usually used to obtain the status information of TiDB for debugging. This document introduces the features of TiDB Control and how to use these features.

Get TiDB Control

You can get TiDB Control by installing it using TiUP or by compiling it from source code.

Note:

It is recommended that the version of the Control tool you use is consistent with the version of the cluster.

Install TiDB Control using TiUP

After installing TiUP, you can use tiup ctl:<cluster-version> tidb command to get and execute TiDB Control.

Compile from source code

  • Compilation environment requirement: Go Version 1.13 or later
  • Compilation procedures: Go to the root directory of the TiDB Control project, use the make command to compile, and generate tidb-ctl.
  • Compilation documentation: you can find the help files in the doc directory; if the help files are lost or you want to update them, use the make doc command to generate the help files.

Usage introduction

This section describes how to use commands, subcommands, options, and flags in tidb-ctl.

  • command: characters without - or --
  • subcommand: characters without - or -- that follow a command
  • option: characters with - or --
  • flag: characters exactly following a command/subcommand or option, passing value to the command/subcommand or option

Usage example: tidb-ctl schema in mysql -n db

  • schema: the command
  • in: the subcommand of schema
  • mysql: the flag of in
  • -n: the option
  • db: the flag of -n

Currently, TiDB Control has the following subcommands:

  • tidb-ctl base64decode: used for BASE64 decoding
  • tidb-ctl decoder: used for KEY decoding
  • tidb-ctl etcd: used for operating etcd
  • tidb-ctl log: used to format the log file to expand the single-line stack information
  • tidb-ctl mvcc: used to get the MVCC information
  • tidb-ctl region: used to get the Region information
  • tidb-ctl schema: used to get the schema information
  • tidb-ctl table: used to get the table information

Get help

Use tidb-ctl -h/--help to get usage information.

TiDB Control consists of multiple layers of commands. You can use -h/--help after each command/subcommand to get its respective usage information.

The following example shows how to obtain the schema information:

Use tidb-ctl schema -h to get usage details. The schema command itself has two subcommands: in and tid.

  • in is used to obtain the table schema of all tables in the database through the database name.
  • tid is used to obtain the table schema by using the unique table_id in the whole database.

Global options

tidb-ctl has the following connection-related global options:

  • --host: TiDB Service address (default 127.0.0.1)
  • --port: TiDB status port (default 10080)
  • --pdhost: PD Service address (default 127.0.0.1)
  • --pdport: PD Service port (default 2379)
  • --ca: The CA file path used for the TLS connection
  • --ssl-key: The key file path used for the TLS connection
  • --ssl-cert: The certificate file path used for the TLS connection

--pdhost and --pdport are mainly used in the etcd subcommand. For example, tidb-ctl etcd ddlinfo. If you do not specify the address and the port, the following default value is used:

  • The default service address of TiDB and PD: 127.0.0.1. The service address must be an IP address.
  • The default service port of TiDB: 10080.
  • The default service port of PD: 2379.

The schema command

The in subcommand

in is used to obtain the table schema of all tables in the database through the database name.

tidb-ctl schema in <database name>

For example, running tidb-ctl schema in mysql returns the following result:

[
    {
        "id": 13,
        "name": {
            "O": "columns_priv",
            "L": "columns_priv"
        },
              ...
        "update_timestamp": 399494726837600268,
        "ShardRowIDBits": 0,
        "Partition": null
    }
]

The result is displayed in the JSON format. (The above output is truncated.)

  • If you want to specify the table name, use tidb-ctl schema in <database> -n <table name> to filter.

    For example, tidb-ctl schema in mysql -n db returns the table schema of the db table in the mysql database:

    {
    "id": 9,
    "name": {
        "O": "db",
        "L": "db"
    },
    ...
    "Partition": null
}

    (The above output is also truncated.)

    If you do not want to use the default TiDB service address and port, use the --host and --port options to configure. For example, tidb-ctl --host 172.16.55.88 --port 8898 schema in mysql -n db.

The tid subcommand

tid is used to obtain the table schema by using the unique table_id in the whole database. You can use the in subcommand to get all table IDs of certain schema and use the tid subcommand to get the detailed table information.

For example, the table ID of mysql.stat_meta is 21. You can use tidb-ctl schema tid -i 21 to obtain the detail of mysql.stat_meta.

{
 "id": 21,
 "name": {
  "O": "stats_meta",
  "L": "stats_meta"
 },
 "charset": "utf8mb4",
 "collate": "utf8mb4_bin",
  ...
}

Like the in subcommand, if you do not want to use the default TiDB service address and status port, use the --host and --port options to specify the host and port.

The base64decode command

base64decode is used to decode base64 data.

tidb-ctl base64decode [base64_data]
tidb-ctl base64decode [db_name.table_name] [base64_data]
tidb-ctl base64decode [table_id] [base64_data]

  1. Execute the following SQL statement to prepare the environment:

    use test;
create table t (a int, b varchar(20),c datetime default current_timestamp , d timestamp default current_timestamp, unique index(a));
insert into t (a,b,c) values(1,"哈哈 hello",NULL);
alter table t add column e varchar(20);

  2. Obtian MVCC data using the HTTP API interface:

    $ curl "http://$IP:10080/mvcc/index/test/t/a/1?a=1"
{
 "info": {
  "writes": [
   {
    "start_ts": 407306449994645510,
    "commit_ts": 407306449994645513,
    "short_value": "AAAAAAAAAAE="    # The unique index a stores the handle id of the corresponding row.
   }
  ]
 }
}%

$ curl "http://$IP:10080/mvcc/key/test/t/1"
{
 "info": {
  "writes": [
   {
    "start_ts": 407306588892692486,
    "commit_ts": 407306588892692489,
    "short_value": "CAIIAggEAhjlk4jlk4ggaGVsbG8IBgAICAmAgIDwjYuu0Rk="  # Row data that handle id is 1.
   }
  ]
 }
}%

  3. Decode handle id (uint64) using `base64decode` .

    $ tidb-ctl base64decode AAAAAAAAAAE=
hex: 0000000000000001
uint64: 1

  4. Decode row data using base64decode.

    $ ./tidb-ctl base64decode test.t CAIIAggEAhjlk4jlk4ggaGVsbG8IBgAICAmAgIDwjYuu0Rk=
a:      1
b:      哈哈 hello
c is NULL
d:      2019-03-28 05:35:30
e not found in data

# if the table id of test.t is 60, you can also use below command to do the same thing.
$ ./tidb-ctl base64decode 60 CAIIAggEAhjlk4jlk4ggaGVsbG8IBgAICAmAgIDwjYuu0Rk=
a:      1
b:      哈哈 hello
c is NULL
d:      2019-03-28 05:35:30
e not found in data

The decoder command

  • The following example shows how to decode the row key, similar to decoding the index key.

    $ ./tidb-ctl decoder "t\x00\x00\x00\x00\x00\x00\x00\x1c_r\x00\x00\x00\x00\x00\x00\x00\xfa"
format: table_row
table_id: -9223372036854775780      table_id: -9223372036854775780
row_id: -9223372036854775558        row_id: -9223372036854775558

  • The following example shows how to decode value.

    $ ./tidb-ctl decoder AhZoZWxsbyB3b3JsZAiAEA==
format: index_value
type: bigint, value: 1024       index_value[0]: {type: bytes, value: hello world}
index_value[1]: {type: bigint, value: 1024}

The etcd command

  • tidb-ctl etcd ddlinfo is used to obtain DDL information.

  • tidb-ctl etcd putkey KEY VALUE is used to add KEY VALUE to etcd (All the KEYs are added to the /tidb/ddl/all_schema_versions/ directory).

    tidb-ctl etcd putkey "foo" "bar"

    In fact, a key-value pair is added to the etcd whose KEY is /tidb/ddl/all_schema_versions/foo and VALUE is bar.

  • tidb-ctl etcd delkey deletes the KEY in etcd. Only those KEYs with the /tidb/ddl/fg/owner/ or /tidb/ddl/all_schema_versions/ prefix can be deleted.

    tidb-ctl etcd delkey "/tidb/ddl/fg/owner/foo"
tidb-ctl etcd delkey "/tidb/ddl/all_schema_versions/bar"

The log command

The stack information for the TiDB error log is in one line format. You could use tidb-ctl log to change its format to multiple lines.

The keyrange command

The keyrange subcommand is used to query the global or table-related key range information, which is output in the hexadecimal form.

  • Execute the tidb-ctl keyrange command to check the global key range information:

    {{< copyable "shell-regular" >}}

    tidb-ctl keyrange
    global ranges:
  meta: (6d, 6e)
  table: (74, 75)

  • Add the --encode option to display encoded keys (in the same format as in TiKV and PD):

    {{< copyable "shell-regular" >}}

    tidb-ctl keyrange --encode
    global ranges:
  meta: (6d00000000000000f8, 6e00000000000000f8)
  table: (7400000000000000f8, 7500000000000000f8)

  • Execute the tidb-ctl keyrange --database={db} --table={tbl} command to check the global and table-related key range information:

    {{< copyable "shell-regular" >}}

    tidb-ctl keyrange --database test --table ttt
    global ranges:
  meta: (6d, 6e)
  table: (74, 75)
table ttt ranges: (NOTE: key range might be changed after DDL)
  table: (74800000000000002f, 748000000000000030)
  table indexes: (74800000000000002f5f69, 74800000000000002f5f72)
    index c2: (74800000000000002f5f698000000000000001, 74800000000000002f5f698000000000000002)
    index c3: (74800000000000002f5f698000000000000002, 74800000000000002f5f698000000000000003)
    index c4: (74800000000000002f5f698000000000000003, 74800000000000002f5f698000000000000004)
  table rows: (74800000000000002f5f72, 748000000000000030)